Shyamashree, I suggest that you compare cases and controls for biomarkers, diseases, age, gender, body mass index, and smoking. As Sharad suggested, you need to find what are your major confonders from a detailed literature search. This sounds like genome tagging and many environmental factors need to be considered. Air quality is important, as in perceived stress and other psychosocial factors such as depression. Consider type of employment and exposure at work to air pollution and chemicals. Remember that skin contact with chemicals is considered an environmental factor. Once you measure all the relevant variables and compare the cases to controls at baseline, then you can decide the inclusion/exclusion criteria. For example, if 2% of the sample worked with chemicals, then this could be an exclusion criteria. In genome studies it is important to have as homogenous a sample as possible. You are comparing an intervention or one specific difference (ethnicity, education...etc) and need to adjust for only a few variables. Another consideration is having a sample large enough for sufficient power to detect a difference (at least 80%) with 95% confidence. The power analysis should take into consideration the number of adjustment variables necessary in the analysis.

you have to provide study details, no one can answer how cases should be different that controls. are you talking about global methylation levels in general or methylation of a specific gene(s). what type of study you are conducting, case-control, crossectional or longitudinal. how do you know that presence of diabetes, htn and lipid lowering drugs are confounding factors. 

Dear Sharad,

I am doing case-control study, and methylation of whole genome by Microarray of CAD patients. Hence, I was wondering about the inclusion and exclusion criteria.

Shyamashree, I suggest that you compare cases and controls for biomarkers, diseases, age, gender, body mass index, and smoking. As Sharad suggested, you need to find what are your major confonders from a detailed literature search. This sounds like genome tagging and many environmental factors need to be considered. Air quality is important, as in perceived stress and other psychosocial factors such as depression. Consider type of employment and exposure at work to air pollution and chemicals. Remember that skin contact with chemicals is considered an environmental factor. Once you measure all the relevant variables and compare the cases to controls at baseline, then you can decide the inclusion/exclusion criteria. For example, if 2% of the sample worked with chemicals, then this could be an exclusion criteria. In genome studies it is important to have as homogenous a sample as possible. You are comparing an intervention or one specific difference (ethnicity, education...etc) and need to adjust for only a few variables. Another consideration is having a sample large enough for sufficient power to detect a difference (at least 80%) with 95% confidence. The power analysis should take into consideration the number of adjustment variables necessary in the analysis.

if your disease if CAD (coronary artery disease) then all the factors you listed in the questions are co-variates. reason is that diabetes, HTN are risk factors for CAD, but reverse is not true. same goes for lipid lowering drugs. 

Confounders are variables which are correlated variables and one can predict other.

Interesting question. Has got a lot to do with levels of folic acid and methyl cobalamine. So diet, vitamin supplements etc will have a lot of implications. Also methyl transferases...their expression levels etc. You may also chose to correlate with serum HCy levels. It has to be determined. Good luck. 

Apurva, excellent observations. Homocysteine levels are important as are  vitamin B, zinc and magnesium through diet and  supplementation. Another chronic disease that is associated with methylation is chronic kidney disease. Genetic considerations are important since there are some polymorphisms that can cause poor functioning of the methylation cycle. Genetic mutations known as SNPs (single nucleotide polymorphisms) within the genes responsible for the production of the needed enzymes can result in a decreased efficiency of these enzymes. This can cause a slowdown of the entire cycle. As previously mentioned, medications can interact with the methlation process.

Shyamashree, vast improvement in microRNAs research and its differential expression pattern in each disease was observed in comparison to controls, which suggests that in hypertension, diabetes and drug intake so on, the pattern of microRNAs expression is different and which is targeting various pathways (methylation pathway to), a study from USA (PMID: 23430482 ) found that miRNAs interaction with homocysteine (Hcy) pathway and involved in epigenetic alterations. Our group also found that one microRNA interacting with one carbon metabolic pathway gene and altering in Hcy levels, in particular diabetes and hypertension we found variation in the particular microRNA expression and alteration the Hcy levels.       

Thank you all for your responses. Like discussed I am looking into the same confounding factors and hope to do some correaletion in the end of it.